From Blind Spots to Gains: Diagnostic-Driven Iterative Training for Large Multimodal Models

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Jia, Hongrui, Jiang, Chaoya, Heng, Yongrui, Zhang, Shikun, Ye, Wei
Format:	Preprint
Published:	2026
Subjects:	Computer Vision and Pattern Recognition
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Internet

https://arxiv.org/abs/2602.22859